Tom Lee Rebuts Enron Warning Over $3T AI Finance Deals

The clash between Burry and Lee reflects a fundamental disagreement about how to evaluate massive off-balance-sheet commitments in the AI era. Burry, whose track record shorting subprime mortgages before 2008 lends him credibility, warns that Nvidia's $500 billion financing pact with a consortium of major financial firms echoes Enron's playbook of hiding leverage outside official accounts. The Wall Street Journal's report that nine tech giants are sitting on $3 trillion in AI obligations—mostly unpaid leases and chip purchase agreements—amplified his concern because the figure towers over Enron's $60 billion bankruptcy. Lee's rebuttal is not that the obligations don't exist, but that they are structurally different from Enron's deception. Off-balance-sheet leases and purchase commitments are transparent, contractually documented, and can be unwound; crucially, they only trigger accounting entries when construction begins, giving companies real exit ramps if market conditions change. He points out that the seven largest tech firms today earn margins among the highest in corporate history—far healthier than the fiber bubble companies of the 1990s that actually invented fake revenue. The deciding factor, both Lee and Burry implicitly agree, is whether the spending drives real productivity gains. Pence noted that the AI buildout already consumes 2% to 2.5% of US GDP and expects it to exceed the Pentagon's $866.6 billion 2026 budget by 2027. Yet adoption remains sparse: only 30% of companies report productivity gains from AI, and just 7% say their rollout is complete. If those numbers climb sharply, the $3 trillion looks like justified infrastructure investment—comparable to transcontinental railroad spending in the 1850s. If they stall, Burry's Enron alarm becomes harder to dismiss.